Table 2.

Sample Size and Study Duration Estimates for a Clinical Trial Using Severe Hypoglycemia as the Primary Outcome^a

SH rate in control groupb	Length of follow-up	SH relative reduction by interventionc
		50%	25%	10%
5	1 month	31,300	145,000	968,000
	3 months	10,400	48,200	323,000
	6 months	5220	24,100	161,000
15	1 month	10,400	48,200	323,000
	3 months	3480	16,100	108,000
	6 months	1740	8030	53,800
30	1 month	5220	24,100	161,000
	3 months	1740	8030	53,800
	6 months	870	4010	26,900
45	1 month	3480	16,100	108,000
	3 months	1160	5350	35,900
	6 months	580	2680	17,900

^aTotal number of subjects. Half would be randomized to each group. Assumes 90% power with type 1 error rate of 5% (two-tailed) and 1:1 allocation to intervention and control. This model likely underestimates the true sample size necessary because it assumes that every subject has the same Poisson rate whereas the true distribution is likely a mixture of Poisson outcomes.

^bNocturnal events per 100 person-years.

^cFor example, a 25% reduction when the SH rate is 30 in the control group would result in a rate of 22.5 in the intervention group.